Package: numatop / 1.0.4-3
Metadata
| Package | Version | Patches format |
|---|---|---|
| numatop | 1.0.4-3 | 3.0 (quilt) |
Patch series
view the series file| Patch | File delta | Description |
|---|---|---|
| 0000 hardening changes for makefile.patch | (download) |
Makefile |
10 5 + 5 - 0 ! |
--- |
| 0001 Use libnuma API to create memory on a specified node.patch | (download) |
test/mgen/Makefile |
2 1 + 1 - 0 ! |
[patch 01/16] use libnuma api to create memory on a specified node Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit With latest Linux kernel, the memory pages gets migrated from the original node to the one where it's now being accessed, then the RMA is changed to LMA. This patch uses libnuma API to tell kernel it doesn't need to migrate the memory pages. Signed-off-by: Jin Yao <yao.jin@linux.intel.com> |
| 0002 Add cache monitoring and memory bandwidth monitoring.patch | (download) |
common/cmd.c |
85 83 + 2 - 0 ! |
[patch 02/16] add cache monitoring and memory bandwidth monitoring features Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit The feature is only supported on Intel Xeon V4 family. Note that this feature requires kernel CQM/MBM patch but the patch has not been upstreamed yet (target for 4.7). So please apply following kernel patch before using numatop cache monitoring and memory bandwidth monitoring features. The link to kernel patch: https://github.com/vshiva1/linux/tree/cqm_rcfix2 Signed-off-by: Jin Yao <yao.jin@linux.intel.com> |
| 0003 Add a feature to show per Node QPI bandwidth.patch | (download) |
common/cmd.c |
14 14 + 0 - 0 ! |
[patch 03/16] add a feature to show per-node qpi bandwidth Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit Signed-off-by: Jin Yao <yao.jin@linux.intel.com> |
| 0004 Add a feature to show per Node memory controller ban.patch | (download) |
common/cmd.c |
6 3 + 3 - 0 ! |
[patch 04/16] add a feature to show per-node memory controller bandwidth Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit Signed-off-by: Jin Yao <yao.jin@linux.intel.com> |
| 0005 common os fix null pointer dereference on nrec.patch | (download) |
common/os/pfwrapper.c |
4 2 + 2 - 0 ! |
[patch 05/16] common/os: fix null pointer dereference on nrec
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 8bit
pf_profiling_record checks that nrec is NULL, however, a subsequent call
from this function to profiling_recbuf_update dereferences nrec by a read
and a write. Fix this by checking if nrec is NULL and returning early
in profiling_recbuf_update to avoid the dereference bug.
Fixes warning from clang scan-build:
common/os/pfwrapper.c:380:6: warning: Dereference of null pointer (loaded from variable 'nrec')
i = *nrec;
^~~~~
Signed-off-by: Colin Ian King <colin.king@canonical.com>
|
| 0006 Do not print garbage latency if lat_data_get fails t.patch | (download) |
common/win.c |
17 11 + 6 - 0 ! |
[patch 06/16] do not print garbage latency if lat_data_get fails to get latency Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit It is possible for lat_data_get to not succeed in reading the latency, in which case the latency contains a garbage value and this misleading value is printed. Instead, print "unknown" if we cannot fetch the latency to indicate that it could not be read rather than emiting an incorrect garbage value. Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0007 Initialise nprocs to zero to avoid printing garbage .patch | (download) |
common/win.c |
2 1 + 1 - 0 ! |
[patch 07/16] initialise nprocs to zero to avoid printing garbage
nprocs value
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 8bit
nprocs is not intialized, so we can call reg_scroll_show passing a
garbage value of nprocs from the stack. Instead, initialize it to
zero so we don't have a random garbage value being emitted.
Fixes clang scan-build warning:
common/win.c:3403:2: warning: Function call argument is an uninitialized value
reg_scroll_show(data_reg, (void *)lines, nprocs,
Signed-off-by: Colin Ian King <colin.king@canonical.com>
|
| 0008 Fix some discarded const qualifiers on pointer targe.patch | (download) |
common/lwp.c |
4 2 + 2 - 0 ! |
[patch 08/16] fix some discarded 'const' qualifiers on pointer target type Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit Building with -Wcast-qual shows that there are several occurrances where const pointer target types are being cast to non-const. Fix this minor issue by adding appropriate constifications. Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0009 Remove duplicate declaration of dump_ctl.patch | (download) |
common/util.c |
1 0 + 1 - 0 ! |
[patch 09/16] remove duplicate declaration of dump_ctl Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0010 Put g_scr_height and g_scr_width into reg.h.patch | (download) |
common/include/page.h |
3 0 + 3 - 0 ! |
[patch 10/16] put g_scr_height and g_scr_width into reg.h Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit g_scr_height and g_scr_width belong to reg.c so it makes sense to add their extern declarations in reg.h and remove the duplication of these from page.h and win.h Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0011 Put g_precise g_ncpus and g_sortkey into numatop.h.patch | (download) |
common/cmd.c |
1 1 + 0 - 0 ! |
[patch 11/16] put g_precise, g_ncpus and g_sortkey into numatop.h Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit g_precise, g_ncpus and g_sortkey belong to numatop.c so it makes sense to add their extern declarations into numatop.h and remove the duplication of these from node.c, os_util.c, win.c Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0012 Make bufdesc_cut a static function.patch | (download) |
common/win.c |
2 1 + 1 - 0 ! |
[patch 12/16] make bufdesc_cut a static function Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit This function is only referenced in win.c, so make it static Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0013 Make llrec_add a static function.patch | (download) |
common/os/os_perf.c |
2 1 + 1 - 0 ! |
[patch 13/16] make llrec_add a static function Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit This function is only referenced in os_perf.c, so make it static Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0014 Make countval_sum a static function.patch | (download) |
common/os/node.c |
2 1 + 1 - 0 ! |
[patch 14/16] make countval_sum a static function Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit This function is only referenced in node.c, so make it static Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0015 Don t compare a double with 0.0.patch | (download) |
common/os/os_util.c |
3 2 + 1 - 0 ! |
[patch 15/16] don't compare a double with 0.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit Compiling with -Wfloat-equal picks up a potential unsafe comparison with 0.0: common/os/os_util.c:228:11: warning: comparing floating point with == or != is unsafe [-Wfloat-equal] instead, compare to see if it is less than a very small value. Signed-off-by: Colin Ian King <colin.king@canonical.com> |
| 0016 Add SKX supporting.patch | (download) |
Makefile |
2 1 + 1 - 0 ! |
[patch 16/16] add skx supporting Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit Signed-off-by: Jin Yao <yao.jin@linux.intel.com> |
